Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: How does Ruffian learn?

Author: Uri Blass

Date: 00:43:01 10/01/02

Go up one level in this thread


On September 30, 2002 at 19:50:24, Andreas Herrmann wrote:

>On September 30, 2002 at 17:15:28, David Dahlem wrote:
>
>>Hi
>>If i understand correctly, Ruffian only learns in Winboard mode, not in UCI
>>mode. Is this book learning only and is the ruffian.bok updated with learning
>>info? Also will Ruffian in UCI mode use the info learned in Winboard mode?
>>
>>Regards
>>Dave
>
>The UCI protocoll doesn't send the game result to the engine, that's the reason
>why engines normaly can't do book learning in UCI mode.

I did not implement book learning in movei but I see no reason to wait for the
result to learn.

The result can be misleading because the program can get a big advantage and
blunder later.

In the last case it seems better not to avoid the line but to have more moves in
book so in the next time the program is going to get the same position with more
time on the clock.

I wonder how authors implement knowledge.
Do they start by thinking how to do it or by reading how other do it?

I think that the right way is to start by thinking.

Uri



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.